U.S. healthcare system's reliance on biased data perpetuates health inequity, report shows

Photo: The Terry Group

There is widespread use of biased utilization data in the U.S. healthcare system, resulting in the misallocation of resources and perpetuation of health inequity, most notably in the already underserved communities nationwide, according to “How Biased Utilization Data Perpetuate Health Inequity: A Two-Part Strategy to Address the Problem,” a new report from The Terry Group, a health and actuarial consulting firm.

The report found the healthcare system’s reliance on incorrect, outdated or incomplete utilization data may not accurately reflect the underlying health risks and needs of disadvantaged populations and communities, resulting in a serious misallocation of healthcare resources.

We sat down with Munzoor Shaikh, executive vice president of customer success, healthcare and Terry Ventures, at The Terry Group, to discuss the findings of the report and understand the firm’s proposed two-part strategy to address the problem.

Q. What is the thrust of your new report? Talk a bit about the problem.

A. Utilization data are employed for a wide variety of purposes in the healthcare system, including establishing provider networks, setting negotiated reimbursement rates and calculating risk adjustment scores. They are also fed into the predictive health models used to target preventive care services to at-risk patients and plan members and to prioritize case management outreach.

The problem is that the data may not accurately reflect the underlying health risks and needs of underserved populations. To begin with, insurance coverage is skewed by socioeconomic status, as well as by race and ethnicity, with Blacks and Hispanics less likely to have coverage than non-Hispanic Whites.

Even when people have coverage, moreover, members of disadvantaged groups in the population may face barriers to access that members of more advantaged groups do not. And when members of disadvantaged groups do interact with the healthcare system, it is more likely to be in the ER, which means that standard health assessment data are often not collected.

Since utilization data play such a central role in steering healthcare financing and delivery, bias in the data can lead to a serious misallocation of healthcare resources that perpetuates health inequity. Nor is this merely a hypothetical concern. A growing number of studies have confirmed that utilization data often significantly underestimate the underlying health needs of minorities.

Q. You have a two-part strategy to fight the problem of biased utilization data. Please describe the first part of the strategy: informed leaps of faith.

A. The first part of the strategy calls for insurers, employers, health systems and provider groups to proactively launch initiatives designed to improve engagement with the healthcare system in at-risk communities. The goal is to begin to offset the adverse impact of biased utilization data on health equity by increasing engagement with the healthcare system.

Launching these initiatives without representative utilization data as a guide will require leaps of faith. However, they need not be blind ones. There are of course national data on the incidence of chronic health conditions by age, gender, and race and ethnicity.

There also is data on social determinants of health variables at the community level. Based on the available data, as well as observation and experience, payers and providers could make informed guesses about those populations and communities where existing utilization data are most likely to understate true needs.

In the report, we discuss a wide range of possible initiatives. Community clinics and mobile health units can bridge gaps in preventive care services. Food as medicine programs can not only help prevent, manage and treat illness, but also provide opportunities for dieticians and case workers to have regular interactions with patients that may suggest the need for additional health interventions.

Nonmedical benefits, such as transportation and childcare services, can remove what are often major barriers to access for underserved populations. Community health training programs can equip community health workers and patient navigators to identify unmet needs and refer people to appropriate medical and nonmedical resources.

Q. Please describe the second part of the strategy: in-depth community studies designed to identify the underlying causes of underutilization and misutilization.

A. While leap-of-faith initiatives can make important contributions to improving health equity, they are a blunt instrument. To refine these initiatives, as well as to make personalized interventions at the individual as opposed to the community level, it is necessary to understand exactly what is driving the underutilization and misutilization of healthcare services.

Knowing how a community ranks in terms of SDOH variables, such as income or the quality of schools and housing, can help inform efforts to improve outreach and engagement. So can knowing the race, ethnicity, and educational attainment of individual patients and plan members.

But ultimately these variables are merely proxies for deeper underlying obstacles to engaging, or appropriately engaging, with the healthcare system. They are correlated with the underutilization and misutilization of healthcare services but are not necessarily the causes.

To understand individual behavior, we must drill down deeper. Perhaps it is lack of trust in the healthcare system that prevents some people from engaging with it. Perhaps it is caregiving responsibilities that prevent others.

Or maybe the reasons lie in language barriers or unmet transportation needs. Such questions cannot be answered, or at least not fully answered, by proxy data on the demographic or socioeconomic group to which people belong or the community in which they live. We call these underlying determinants of individual behavior causative engagement drivers to distinguish them from proxy variables.

Understanding these causative drivers is where the in-depth community studies we propose come in. As we envision them, these studies would begin with an exploratory phase in which information and data related to the lived experiences of people in the community are gathered through extensive observation and interviews with key community actors.

Based on what is learned, a survey would then be conducted in the community with the goal of identifying “sustainable pathways” through which patient behavior and patient outcomes can be influenced.

The insights gained from these community studies could improve health equity in several ways. They could help in refining the leap-of-faith initiatives that were already launched in the first part of the strategy, as well as in designing new ones. They could also help retrain the algorithms used to predict health risks and needs.

The greatest benefits, however, would likely come from using the insights to personalize patient outreach and engagement. Once the survey has been completed and analyzed, individual patients and plan members would also be asked to answer those questions that yielded the most valuable information about obstacles to engagement in the community, and their responses would be integrated into their medical records.

This could be done during plan enrollment and annual wellness visits or through special outreach programs.

We believe that in-depth community studies of this kind have the potential to greatly improve engagement with the healthcare system in at-risk communities, increasing appropriate utilization and reducing inappropriate utilization.

Over time, the utilization data on which healthcare system participants rely would become more representative, leading to a more equitable allocation of overall healthcare resources. And as the allocation of healthcare resources improves, so will health equity.

While the goal of the first part of the strategy is to offset the adverse impact of biased utilization data on health equity, the goal of the second part is to eliminate it by gradually improving the utilization data itself.

Needless to say, this will not happen overnight. A successful study will need to be an iterative process in which utilization patterns are analyzed, observational and survey data are collected, causative engagement drivers are identified and integrated into patient records, stepped-up patient outreach and engagement occur, the impact on utilization patterns is analyzed, and the whole process is repeated.

While a successful study will require a significant commitment of time and money, the results could be life-changing for underserved populations and communities.

Along with the health benefits for individuals, there would also be financial benefits for payers and providers. The potential to realize significant returns on investments in health equity is obviously greatest in value-based payment arrangements, where improvements in health translate directly into improvements in the bottom line.

But there are also significant opportunities to realize positive returns in fee-for-service payment arrangements.

Q. What should healthcare provider organization CIOs and other health IT leaders be doing today to help fight the problem of biased utilization data?

A. CIOs and other health IT professionals have access to a great deal of day-to-day operational data, including website/call center usage statistics and encounter statistics in different treatment settings, most of which can be cross-tabbed with patient/member demographics.

Just like the utilization data used in allocating healthcare resources, this operational data may fail to reflect underlying health risks and needs due to underutilization and misutilization of healthcare services by disadvantaged populations. IT leaders can help to promote health equity by identifying patterns in the data they use that may indicate an underlying bias.

There may be a problem, for instance, when operational data seems poorly aligned with what one would expect to see based on other data sources. Sometimes these other data sources will be internal.

For example, IT leaders may notice that the ethnic diversity in their patient/member demographics is not reflected in their encounter data. Although IT professionals cannot solve the problem themselves, they can flag it so that other departments in their organization can then devise appropriate solutions, such as providing transportation or translation services.

Sometimes the other data sources will be external ones that IT leaders routinely consult. As an example, encounter data in a given geography may not correlate with social vulnerability index data, suggesting that there may be gaps in insurance coverage, barriers to access, or member selection bias. Once again, flagging the potential problem can lead to solutions.

Another thing to look out for is rapid shifts in utilization patterns. Since the COVID-19 pandemic began, our healthcare system has suffered many shocks. In 2020, for example, the number of in-person healthcare visits plunged as lockdowns and social distancing disrupted routine medical care.

Now, in 2022, the number of ER visits is spiking. Are the two developments related? Probably. Based on the first, could the second have been predicted? Again, probably.

IT leaders are uniquely positioned to track shifts in utilization patterns in real time. Catching them early not only gives healthcare organizations time to adjust and prepare, but can also improve health equity. Just as the pandemic brought to light latent inequities in our healthcare system, so can rapid shifts in utilization patterns.

Whether it’s a decline in wellness visits or a spike in ER visits, they are a sign of underlying problems that need to be addressed.

Twitter: @SiwickiHealthIT
Email the writer: [email protected]
Healthcare IT News is a HIMSS Media publication.

Source: Read Full Article